AI Paper Atlas

Research map / Optimization / Implicit bias and generalization

On the Interpretability of Regularisation for Neural Networks Through Model Gradient Similarity

Vincent Szolnoky, Viktor Andersson, Balazs Kulcsar, Rebecka Jörnsten

NeurIPS 2022 · Optimization › Implicit bias and generalization · approach: Explicit and implicit regularization interplay · 2 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add MGS trace penaltyMGS · MNIST LeNet-5 80% noisetest accuracy: 23.9 -> 74.5Improved
add MGS penalty vs dropoutMGS · MNIST LeNet-5 80% noisetest accuracy: 39.5 -> 74.5Improved
add MGS penaltyMGS · MNIST LeNet-5 0% noisetest accuracy: 89.6 -> 95.2Improved
add MGS penaltyMGS · MNIST FCN 80% noisetest accuracy: 24.8 -> 51.4Improved
add dropoutbaseline · MNIST FCN 0% noisetest accuracy: 82.1 -> 79.5Hurt
add MGS penaltyMGS · Facial Keypoints LeNet-5 noise 0test loss: 68.2 -> 14.1Improved
add MGS penalty vs dropoutMGS · Facial Keypoints LeNet-5 noise 30test loss: 19.7 -> 17.8Improved
add MGS penaltyMGS · Facial Keypoints FCN noise 0test loss: 281.1 -> 10.6Improved
add dropoutbaseline · Facial Keypoints FCN noise 0test loss: 281.1 -> 208.2Mixed
change learning rate in training-parameter robustness testMGS · MGS on corrupted MNISTtest accuracyHurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy